The Patient Protection and Affordable Care Act (ACA) went into effect in the U.S. in March 2010, and phased implementation is scheduled through January 2018. The need for reform is clear: health care spending per capita in the U.S. is the highest in the world. Despite this, health care quality and life expectancy are worse in the U.S. compared to similar countries. Patients have low satisfaction with the care they receive, and nearly 30 million people still are without health insurance (1,2,3).
An important goal of the ACA is to improve health care quality and reign in escalating costs by linking the two in a new model of health care delivery and payment called Accountable Care Organizations (ACOs). ACOs provide care to a defined patient population with shared financial incentives and risk between providers and payors. ACOs may be formed by physicians contracting with hospitals, hospitals that employ physicians or joint ventures between hospitals and physician groups (1,2). Simultaneously, exemplified by the Centers for Medicare & Medicaid (CMS) Hospital Readmission Reduction Program (HRRP), both government and private insurers are beginning to implement “pay for performance” paradigms. The HHRP focuses specifically on hospital readmission as a quality metric and imposes financial penalties at institutions with high readmission rates for specific medical conditions (4). How does practicing within an ACO and evolving pay for performance model impact pediatric neurosurgery?
First of Its Kind
Primary Children’s Hospital (PCH) is a tertiary pediatric hospital that is part of Intermountain Healthcare but is staffed by pediatric surgical specialists who are University of Utah faculty members and are employed by the University of Utah Medical Group (UUMG). In 2015, PCH and UUMG came together to create an ACO called Pediatric Specialty Services (PSS), which contracts exclusively for subspecialty care for 77,000 children in Utah, commercially insured via SelectHealth. Physicians are reimbursed according to fee-for-service contracts via a capitated advanced payment model based on five-year historical data with annual reconciliation. If health-care claims are less than the capitated budget, SelectHealth pays the difference to PSS, with revenues shared equally between PCH and UUMG. In the first year of implementation, PSS was associated with a 20 percent decrease in hospital admission for children presenting to the emergency department with asthma, without increasing the number of repeat visits. Astoundingly, 65 percent of the monthly variation in health-care costs of the population of 77,000 children was due to only three high-cost inpatients! In 2016, PSS was expanded to contract for the care of an additional 57,000 children in Utah insured via Community Care, the Medicaid product of SelectHealth.
The PSS has not affected the delivery of pediatric neurosurgical care to date. This is not surprising as surgical care has not been a focus in early ACOs, and surgeon leadership has been lacking (5,6). While the PSS is unique as the first subspecialty only ACO, it is sure to be emulated by other health care organizations. The PSS model demonstrates that subspecialty care, including neurosurgery, is an increasingly important focus of the evolving ACO and HRRP pay-for-performance model. Readmission rates following surgery vary widely, from 3.8 to 41.0 percent in published literature (4). Unplanned hospital readmissions are estimated to cost greater than $17 billion annually (7).
The HHRP recently expanded pay-for-performance efforts to link readmission rates to payment for surgical procedures in addition to medical conditions beginning with hip and knee arthroplasty. Both government and private insurers have started using commercially available 3M-PPRs software to identify “potentially preventable readmissions” (PPRs) through billing databases and imposing financial penalties for readmission rates that exceed a specific threshold. At pediatric hospitals, readmission following an index CSF shunt operation is among the most common and the single most expensive PPR. Further, 3M-PPR software estimates that 86 percent of readmissions after a CSF shunt procedure are potentially preventable (8,9).
Defining the Factors for Quality and Payment
ACOs, like the PSS, provide an opportunity for neurosurgeons to understand variation in the cost of care we provide and define clinically meaningful performance metrics linked to quality neurosurgical care for our own practice and patients. Is hospital readmission a reasonable surrogate measure of quality in neurosurgery? Are the factors that influence readmission similar between adult and pediatric surgical care? Can appropriate metrics of surgical quality that influence reimbursement be derived from coded administrative databases designed for billing? The surgical literature has recently begun grappling with critical questions like these. Surgeon leadership is essential as the mechanics of how quality and payment are linked have significant implications for neurosurgery.
In 2015, surgeons at Johns Hopkins performed a single-center retrospective study to evaluate the etiology of variation in 30-day readmission rates at their hospital over a five-year study period across eight subspecialties, including 2,975 readmissions for 22,559 patients (13.2 percent 30-day readmission rate) (4). They found a variation between subspecialties of 2.1 (breast, endocrine or melanoma surgery) to 24.8 percent (transplant surgery). Over 82 percent of the variation was due to patient-specific factors, including the number and severity of preoperative comorbidities. Surgical subspecialty accounted for 14.5 percent of the variation in readmission and surgeon-specific factors accounted for just 2.8 percent. However, 30 percent of the patients readmitted had at least one post-operative complication (4).
The authors concluded that readmission rates are driven primarily by patient factors that are out of the control of the surgeon. Neurosurgery and pediatric patients were both excluded from this study. However, studies in children suggest that reducing the rate of unplanned hospital readmission in children may offer a significantly smaller cost savings opportunity compared to adults (8). Further, studies of complications in pediatric neurosurgery suggest only a weak link between patient comorbidities and surgical complications (10).
In 2013, neurosurgeons at the University of California, San Francisco (UCSF) performed a single-center retrospective review of 30-day readmissions following spine surgery over a 3 1/2 year period (7). This included 5,780 surgeries by 20 surgeons, with a 4.9 percent, 30-day readmission rate. They compared billing datasets identifying readmission for any cause with clinical databases to determine the reason for readmission. Among these, 39.8 percent were due to infection, 13.4 percent were for non-surgical management (primarily inadequate pain control) and 13.9 percent were for a planned staged procedure. They concluded that 25 percent of readmissions were not related to spine surgery, and administrative databases are unable to accurately determine clinically meaningful readmission rates that may be surrogates of quality in spine surgery (7).
Studies of readmissions following cerebrospinal fluid (CSF) shunt failure in pediatric neurosurgery have drawn similar conclusions (9,11). In 2015, neurosurgeons at Children’s Hospital of Atlanta performed a four-year retrospective review of patients readmitted within 30 days after CSF shunt placement or revision. They determined that 21 to 42 percent of readmissions were potentially preventable, depending on whether the readmission of patients with excellent radiographic shunt placement was considered preventable. Although surgical complications were significantly associated with preventable readmissions, they concluded the PPR for patients undergoing CSF shunt surgery was considerably less than the 86 percent estimated by 3M-PPR software using administrative databases (9).
In 2016, the pediatric neurosurgery group in Memphis performed a retrospective review of CSF shunt failure within 90 days at their institution (11). They identified a 21.8 percent, 90-day shunt failure rate, with almost 70 percent of cases occurring within the first 30 days after surgery and a higher rate of shunt failure after revision compared to new shunt placement. They determined that only one third of shunt failures within 90 days were preventable, with approximately half of those cases due to malpositioned ventricular catheters. The authors proposed the “preventable shunt revision rate” as a quality measure in pediatric neurosurgery (11).
While the link between quality neurosurgical care and all-cause hospital readmission is opaque, it is a metric already being tied to reimbursement in the ACO and pay for performance era. The link between clinical outcomes and reimbursement established by the ACA and embodied by ACOs is growing stronger, and it is critical that neurosurgeons lead the way in defining clinically meaningful metrics of quality for patients requiring neurosurgical care.
References
1. Stain SC, Hoyt DB, Hunter JG, Joyce G, Hiatt JR (2014). American surgery and the Affordable Care Act. JAMA Surg 149:984-985.
[aans_authors]
